# File for the paper 
# NONPARAMETRIC INSTRUMENTAL VARIABLE METHODS FOR DYNAMIC TREATMENT EVALUATION
# Gerard J. Van den Berg, Petyo Bonev, Enno Mammen
# Corresponding author: petyo.bonev@unisg.ch
# STRUCTURE OF THIS FILE
# 1. MASTER FILE OF CODES: WHICH RESULTS BY WHICH CODES IN WHICH FILES
# 1.1 Results in main paper
# 1.2 Results in Appendix
# 2. HOW TO GET ACCESS TO THE DATA: A DETAILED DESCRIPTION
# 3. DATA DICTIONARY
# 4. WHICH SOFTWARE
##################
# 1.1. RESULTS IN MAIN PAPER:
# File Data_preparation.txt  - construct variables and extract the sample. Results in datasets Treated.csv, Untreated.csv 
# File MAIN_RESULTS.txt: produces estimates of the TE with confidence bounds, as well the Naive TE. The results are plotted in figure 3a.
# File Testing_for_no_anbticipation.txt: produces figure 3b.
##################
# 1.2 RESULTS IN THE APPENDIX
# 1.2.2 Model diagnostics
# Appendix B.3.1 (i): tests for equality of pretreatment characteristics
# File: Descriptives_and_assumptions_tests.txt: descriptive stats and tests for equality of pretreatment characteristics (exception: pre-jobloss wages). Also creates figure 2 in the appendix.
# File: Test_equality_prejobloss_wages.txt: tests for equality of prejobloss wages
# File: Second_control_group.txt: selects a second control group and performs estimation with this group, as well as a comparison of pretreatment characteristics. Produces figure 3 in the appendix. 
# Appendix B.3.3: Simulation study code
# file Simulation_Study.pdf contains the code for the simulations study. 
#################################################
#################################################
# HOW TO GET ACCESS TO THE DATA: A DETAILED DESCRIPTION
# A. STEPS
# B. CONTACT PERSONS

# A. STEPS
1. Apply for access to the data. The necessary form can be downloaded under:
https://www.comite-du-secret.fr
This requires a statement of the purposes of access to the data Example: replication of an existing project. The code name for our project is REFPARE.

2. After a positive decision of the committee, a data security statement has to be sent. All necessary documents are sent in this case by the data security committee (CASD).

3. The next step is signing a contract with CASD for access to the data. Upon settling the contract, an electronic box is sent to the researcher. The data access is given remotely through this technology. An obligatory intermediate step is a one workshop in Paris on data access and rules for working with sensitive data. 

#B. CONTACT PERSONS 
Who is your contact person: 
1. Initial contact person is the Comite Du Secret Statistique (CASD), 
email: comite-secret@cnis.fr. 
This is the committee responsible for step 1. 

2. After going through step 2, all technical and administrative questions are handled by the service team of the 

Centre D'Acc�ss S�curis� Aux Donn�es (Secure Data Access Center)
Groupe des �coles nationales d��conomie et statistique (GENES) 5,
avenue Henry Le Chatelier - TSA 16643
91764 Palaiseau cedex
website: https://www.casd.eu/en/
email:service@casd.eu 

################################################
################################################
# 3. DATA DICTIONARY: all variables are defined in section B.4 in the online appendix of the paper


################################################
################################################
# 4. WHICH SOFTWARE:
The initial datasets are in SAS-format. Any SAS-version can be used to transform them into CSV. format.
The data preparation and subsequent estimation procedures are performed with R. Any version after R.2.14.1 (Dec 2011) can be used. 



